The $ 300 billion market: Who is the driver and who is the passenger in the self-driving car and how to insure them

How safe are self-driving cars? This question has been facing automakers for about ten years, but insurers can be the first to find the answer.

In the absence of an answer to this question, there is a chasm between automakers / tech companies looking to build self-driving cars, the market and consumers skeptical about the safety of self-driving vehicles (which does not have a human driver).

Building a safe self-driving vehicle is critically necessary for perfect software, quality hardware and thoughtful integration – all these factors together will prove that the autopilot system can function safely. Safety standards such as ISO 26262, SOTIF and ANSI / UL 4600 will help achieve all of these aspects. In particular, UL 4600, published earlier this year, should provide a standardization of the safety case for various systems. UL 4600 should help developers explain and prove the safety of their vehicles (even before these vehicles go into series production).

Stakeholders include federal regulators, municipal authorities, insurance and industry companies, and consumers. And it is insurance companies that can help build consumer confidence in self-driving vehicles. People may not like insurers, but they tend to trust their cold-blooded risk assessments.

An example is JD Power survey in fall 2018… Auto insurance clients were asked who they trust most in the context of self-driving safety testing. 500 people (34% of the total surveyed) said they trust the Insurance Institute for Highway Safety and its subsidiaries. Automakers (12%), the federal government (9%), and state governments (4%) lagged far behind.

Who is driving?

Insurance companies know a lot about the average driver these days. But in the case of self-driving cars, another question arises – who is considered a driver? In the context of self-driving, insurers must consider the driver and evaluate the software installed in the robotic vehicle, Michael Wagner, co-founder and CEO of Edge Case Research, said in a recent interview with the EE Times.

In his blog on Medium Wagner reportedthat Edge Case is bringing Risk Management Services to its business. His company is developing in conjunction with Liberty Mutual.

Given the fact that autonomous vehicles will soon cease to be a purely scientific project, insurance companies must have sufficient data to be confident in insuring the risks associated with autonomous driving.

In the case of conventional vehicles, insurers assess the risks of applicants by introducing their personal data and data from their archives (and they are very large for insurance companies) into algorithms that assess various factors using coefficients. To take into account all the risks, companies also pay attention to rating factors – this allows predicting the likelihood of insurance compensation. The rating allows insurance companies to calculate a price based on the projected cost of taking financial responsibility.

However, unmanned driving will rely entirely on artificial intelligence. Where is the data that insurers can use in assessing the risks associated with a cyber driver replacing a human driver?

There are very few such data to date.

UL 4600

Some insurance companies believe that the first step should be to participate in standards development teams such as UL 4600.

Rather than waiting for the data they need, insurance companies choose to stay ahead of everyone else. Liberty Mutual, AXA and Munich Re America are all of these companies that have already joined the UL 4600 technical standardization group.

At the AVS 2020 virtual panel, Benjamin Lewis, Director of Strategic Partnerships for Automotive and Mobility at Liberty Mutual Insurance, made the following statement: “As a member of the technical standards group, we were involved in the development of UL 4600, the first standard specifically addressing systems security. unmanned driving “. He called the move “an important effort to significantly advance automotive safety innovation.”

In essence, Liberty Mutual needs the knowledge to assess risks and work with partners to manage and mitigate risks.

Importance of Hologram

However, the collaboration between Liberty Mutual and Edge Case Research goes beyond work on UL 4600.

Wagner explained that their company is now involved in software, tools and technical consulting, and through the partnership with Liberty Mutual, they will also be able to provide risk management services.

Edge Case Research is not prepared to fully disclose the details of its risk management plans. One thing’s for sure, though – effectively managing the risks associated with autonomous vehicles is a top priority for the insurance industry. The $ 308.8 billion auto insurance market is at stake – this is IBISWorld assessmentbased on revenue in 2020.

The partnership between Liberty Mutual and Edge Case Research began in June 2019, when Liberty Mutual Strategic Ventures (the venture capital arm) took part in the seed funding of Edge Case Research.

When cooperation announcement Liberty Mutual named “the pilot version Hologram by Edge Case Research»A promising technology that Liberty will invest in to enable the product to grow. The press release described the software as a tool that “tests the hardware and software used by unmanned vehicles to recognize the environment.”

Hologram automatically recognizes scenarios in which self-driving vehicles cannot detect pedestrians, vehicles and other important road users. Those objects that were not detected are marked with a red frame. A continuous risk analysis platform enables auto engineers to learn about security threats before accidents happen (Source: Edge Case Research).

Edge Case Research does not underestimate the importance of modeling. However, Wagner noted that autonomous vehicle fleets could take years to accumulate the billions of miles needed by automakers to demonstrate the safety of their systems. Wagner stated that modeling is critical, but it is also true that “modeling cannot reveal all unknown variables to us.”

This is where Hologram comes in. This software is a tool for analyzing the response of neural networks to real-world data. Of course, such an analysis can be done by comparing what neural networks see with prepared labels. The problem is that this approach scales very poorly – there are too many objects to be tagged and compared.

On the other hand, Wagner says the Hologram is like a kind of “vision test.” Edge Case offers a platform that provides “smart test of perceptual algorithms against conflicting patterns.” Self-driving vehicle companies can analyze data from their sensors, and Hologram can help identify risks. Hologram provides developers with the information they need to retrain algorithms and improve their reliability.

Edge Case Research says it is working with the insurance industry to quantify risk using Hologram. (Source: Edge Case Research)

Impact of unmanned vehicles on insurance companies

The dynamics of the relationship between self-driving car manufacturers and insurance companies was discussed in the Harvard Business Review articlepublished in December 2017.

The article cited the following warning:

Since insurers are in the business of insuring personal cars, they have every reason to be concerned about the growth and profitability of this market. If the number of private owners decreases, then insurance premiums will also decrease. Since before 94% of accidents are related to the human factor, the number and severity of accidents and insured events will decrease, which will also lead to lower insurance premiums, as insurers set rates in accordance with real risk.

This text also identifies three areas that can compensate insurers for lower earnings. These include cybersecurity (the market is estimated at $ 12 billion in annual premiums and grants), the huge liability of product companies (the risks of manufacturers associated with potential software bugs, memory overflows and algorithm defects – about $ 2.5 billion a year) and infrastructure insurance ( cloud servers, signal processing and other safety systems to protect road users – about $ 500 million in premiums).

How insurance companies will interact with unmanned vehicles is not yet clear. Insurers will be able to clarify this issue if: they collect the basic knowledge necessary for “reliable risk management” and actuarial information that insurance analysts can apply in mathematical methods of risk assessment.

Edge Case Research hopes to provide Liberty Mutual with both. The bottom line is that in order for self-driving cars to become a reality, “the market needs to agree,” Wagner said. Edge Case Research believes it can develop the tools insurers need to properly assess risk.

• Russia’s first serial control system for a dual-fuel engine with functional separation of controllers
• In a modern car, there are more lines of code than …
• Free Online Courses in Automotive, Aerospace, Robotics and Engineering (50+)
• McKinsey: rethinking electronics software and architecture in automotive